Submitted to: Interbull Annual Meeting Proceedings
Publication Type: Proceedings
Publication Acceptance Date: September 5, 1997
Publication Date: N/A
Interpretive Summary: Genetic evaluations can be more accurate if yields on test day are analyzed instead of lactation yields. One method being investigated for the direct use of data from test days is a multiple-trait approach with different test days treated as different traits. As part of this approach, information on how the test-day yields are related to one another over the course of lactation is being investigated. Components of variation for milk, fat, and protein yields during first lactation were determined from test day data from 17,190 cows representing 37 large herds in Pennsylvania and Wisconsin and extended to the entire range of the lactation. Test-day information for milk and protein yields was found to be more highly correlated than for milk and fat yields. Development and implementation of a multiple-trait, test-day evaluation system will provide dairy breeders with genetic evaluations that allow more accurate selection for desired traits.
(Co)variance components for milk, fat, and protein yields during first lactation were calculated from test-day data from 17,190 cows representing 37 large herds in Pennsylvania and Wisconsin. Initially, four lactation stages of 75 days each were defined, and the test day nearest the midpoint of each interval was used. This approach minimized the number of lactations with missing values. A canonical transformation was used to estimate variance components, which required that all observations with missing values be deleted. Preliminary analysis showed little effect from this selection. Heritabilities usually increased with lactation stage, were highest for milk, and averaged .15. Phenotypic and genetic corre- lations between milk and protein yields were higher than between milk and fat yields. For each yield trait, the genetic correlation declined from about .90 for adjacent lactation stages to about .75 between lactation stages 1 and 4. When yield traits were from the same lactation stage, the genetic correlation averaged .39 between milk and fat, .78 between milk and protein, and .53 between fat and protein. Phenotypic correlations within lactation stage were >.90 between milk and protein and around .65 between milk and fat and between fat and protein. Estimates for the four 75-day lactation stages were extended to provide estimates for twelve 25- day stages (36 traits) using (co)variance functions, which allowed denser coverage of the lactation.